Ethical considerations in AI: In defence of uncertainty

Abeba Birhane

Imagine a world where we are able to predict people’s behaviour with precision. A world, for example, where we can tell whether someone is going to a commit crime before they do. Many of our problems would just disappear. The quest for absolute certainty has been at the top of Western science’s agenda.

In a similar fashion, current technological developments tend to strive for generalizability and predictability. We value certainty, stability and uniformity. Whereas most of reality, instead of being orderly and stable is seething with change, disorder and process. People, far from being predictable and predetermined, are complex, social and dynamical beings that inherently exist in a web of relations.

This talk discusses how absolute certainty is not only an unattainable goal so far as understanding people and the social world is concerned but also part of the reason we are currently experiencing algorithmic injustice and bias.


Track 2
Location: Gielgud Date: September 30, 2019 Time: 5:05 pm - 5:50 pm Abeba Birhane Abeba Birhane